JOURNAL OF RESEARCH IN SCIENCE TEACHING VOL. 40, NO. 8, PP. 757–775 (2003)Developing the Changes in Attitude about the Relevance of Science (CARS)Questionnaire and Assessing Two High School Science ClassesMarcelle A. Siegel,1Michael A. Ranney21Lawrence Hall of Science, #5200, University of California, Berkeley,California 94720-52002Graduate School of Education, University of California, Berkeley, California 94720-16 70Received 28 January 2002; Accepted 12 November 2002Abstract: This study has two purposes: (a) methodological—to design and test a new instrument ableto reflect changes in attitudes toward science over time, and (b) investigative—to find out the effect of twosimilar curricular treatments on the attitudes of two classes. Items about the relevance of science tostudents’ lives were developed, pilot-tested, and analyzed using Rasch modeling. We then divided reliableitems into three equivalent questionnaire forms. The final three forms of the questionnaire were used toassess high school students’ attitudes. Over 18 weeks, one class used a core curriculum (Science andSustainability) to learn science in the context of making decisions about societal issues. A second classused the same core curriculum, but with parts replaced by computer-based activities (Convince Me)designed to enhance the coherence of students’ arguments. Using traditional and Rasch modelingtechniques, we assessed the degrees to which such instructional activities promoted students’ beliefs thatscience is relevant to them. Both classes tended to agree more, over time, that science is relevant to theirlives, and the increases were statistically equivalent between classes. This study suggests that, by usinginnovative, issue-based activities, it is possible to enhance students’ attitudes about the relevance of science.ß 2003 Wiley Periodicals, Inc. J Res Sci Teach 40: 757–775, 2003‘‘Rose-colored glasses.’’ ‘‘Half full or half empty?’’ Such sayings remind us of the effect thatone’s attitude can have on one’s experience (Ranney, 1996). In the educational and psychologicalresearch literature, many ways of categorizing attitudes have been developed, from disposition toopinion, and from affect to belief. Generally, the research on dispositions tends to examine humansensibilities and approaches; the work on opinions tends to include individuals’ current attitudesor views of a particular issue; research on affect deals with emotional expression and feelings; andthe belief literature concerns deeply held doctrines, such as epistemological ideas about the natureof knowledge.Contract grant sponsor: NSF Research Training Grant.Correspondence to: M.A. Siegel; E-mail: [email protected] 10.1002/tea.10110Published online in Wiley InterScience (www.interscience.wiley.com).ß 2003 Wiley Periodicals, Inc.Attitudes toward ScienceWork in the realm of students’ attitudes toward science has been motivated by the desire toincrease interest, performance, and student retention in science (Third International Mathematicsand Science Study, 2001). Educational studies have produced mixed results but tend to showthat attitudes affect students’ persistence and performance (for a review, see Schommer, 1994).Modest positive correlations between science attitude and science achievement have beenreported in many studies (Schibeci & Riley, 1986; Keeves & Morganstern, 1992). Modelspropose that science instruction influences attitudes in ways that predict achievement (Hegarty-Hazel, 1990; Schibeci & Riley, 1986; Simpson & Oliver, 1990). In particular, science in-struction that is activity-based (Dickinson, 1976; Fraser, 1980; Freedman, 1997) and issue-oriented (Iskandar, 1991; McComas, 1993) has been shown to enhance positive attitudes towardscience.Another reason to extend research on scientific attitudes is that attitudes and beliefs are part of‘‘cognition’’ itself (Schoenfeld, 1985). Research has suggested that people who view science in asophisticated way are better able than others to use their knowledge in more contexts and to makesense out of complex information (Davis, 1998; King & Kitchener, 1994; Linn & Songer, 1993).Thus, science attitude research may be viewed as fundamental to understanding scientificcognition.Measuring Attitudes toward ScienceTo measure attitudes toward science, researchers have used Likert scales, Thurstone scales,semantic differential scales, direct interviews, and, much less frequently, indirect or projectivemethods and interviewing techniques (Ramsden, 1998). The most common method, Likert scales,has several advantages: They are not difficult to create, can include a large number of items thatcan be answered quickly, can provide precise information about a respondent’s degree ofagreement or disagreement, and can provide high reliability (Oppenheim, 1992). Disadvantagesof Likert scales can sometimes be overcome using special techniques or multiple methods.For example, one criticism of measuring attitudes with Likert scales is that the neutral midpointmight not represent neutrality in the subject’s mind, but rather a state of confusion or mis-understanding. Researchers can mitigate this problem by adding another response choice, suchas ‘‘do not understand.’’ Another criticism of Likert scales is that they might decontextualizethe issue studied. For example, students’ diverse views of science might hinge on what is meantby science, and a scale might not be able to account for these different meanings (Bell & Linn,2002). This type of problem is lessened by crafting Likert items to address the specific subtletiesof issues (Oppenheim, 1992) and by using multiple methods. Interviewing and using newopen-ended instruments such as the Views of Nature of Science questionnaire (Lederman,Abd-El-Khalick, Bell, & Schwartz, 2002) offers practical solutions that can add qualitativedepth to a study—although they also increase the classroom time used. Whereas attitude scalesentail problematic assumptions owing to the ways participants might interpret a given item(Lederman et al., 2002), open-ended instruments entail similarly problematic assumptionsregarding how to interpret participants’ statements. Projective or indirect techniques allow theresearcher to uncover ideas without participants’ direct awareness; however, such techniqueshave disadvantages as well, such as lower reliabilities and more interpretive difficulties(Oppenheim, 1992).To our